The author makes a significant error which falsies his entire line of reasoning. The number of continuous variables in a 1000 qubit register is 2000, not 2^1000. Furthermore, the least technically difficult application of a qubit is to create an ALU operating on two qubit based registers. In that system the only entanglement is between the two electrons in a cubit, something that has been accomplished. The number of discrete states which can be held by each qubit depends on how noise free the system is, which is where much focus is at now. The advantage of this system over a conventional computer is that only a small number of qubits are needed to store the integer values of a 1000 bit conventional register. As a consequence, factorization of the large primes in a 1000 bit encryption key can be accomplished more efficiently. Presently, you need a BigNum like representation in conventional computers, which is primarily a software process. The real value of a qubit ALU is more aparent in factorization of the 2048 and 4096 bit keys of RSA or other public key encryption systems. The integer value represented by a 4096 bit key requires 64 registers on a 64-bit machine. Most 64 bit machines do not have 64 registers available for extended mathematical operations and so most of the work must be accomplished in software, and as a consequence it is very slow.
